Math Never Seen

TUGboat, Volume 31 (2010), No. 2 221 Math never seen ò Quality criteria Johannes Küster What makes a notation superior to another? What makes a symbol successful (in the sense that other mathemati- Abstract cians accept and adopt it)? e following list gives the Why have certain mathematical symbols and notations most important quality criteria. A mathematical symbol gained general acceptance while others fell into oblivion? or notation should be: To answer this question I present quality criteria for • readable, clear and simple mathematical symbols. I show many unknown, little- • needed known or little-used notations, some of which deserve • international (or derived from Latin) much wider use. • mnemonic I also show some new symbols and some ideas for new notations, especially for some well-known concepts • writable which lack a good notation (Stirling numbers, greatest • pronounceable common divisor and least common multiple). • similar and consistent • distinct and unambiguous • adaptable Ô Introduction • available For TEX’s "20th birthday it seems appropriate to present is list is certainly not exhaustive, but these are the most some ne points of mathematical typography and some important points. Not all criteria are equally important, ideas for new symbols and notations. Let’s start with a and some may conict with others, so few symbols really quotation from e METAFONTbook [ , p. ]: fulll all criteria. — Let me explain each point in turn. Above all, a notation should be readable — but what “Now that authors have for the rst time the power constitutes readability? Certainly it comprises clear and to invent new symbols with great ease, and to have simple. Also a notation should be short, at least it should those characters printed in their manuscripts on a make an expression shorter than writing out the same state- wide variety of typesetting devices, we must face ment with words. Some of the other criteria contribute to the question of how much experimentation is de- readability as well. sirable. Will font freaks abuse this toy by overdo- When a good, widely accepted notation already exists, ing it? Is it wise to introduce new symbols by the there is no need to invent a new one. So a new notation thousands?” should be needed or necessary. Most mathematical symbols are international (even We all know that METAFONT didn’t become widely ac- if they are given dierent names in dierent languages cepted. But even with other font editors, font freaks did and although there are dierent traditions in mathemati- not create new symbols by the thousands. So while maybe cal notation, e.g. the use of a dot or a comma as decimal METAFONT was too complicated, and its way of thinking separator). Of course a new notation should be interna- foreign to most designers, this can’t be the real reason tional. In the case of an abbreviation (like “sin”, “log”, why only very few new symbols showed up. In fact, to etc.), it should be derived from Latin, as most scientic design a new useful symbol is by no means an easy task, terminology stems from Latin (and Greek), and so does which I hope will become clear in the following. Just as we the international vocabulary of mathematics. all do a lot more reading than writing, it is much easier to A notation should be easy to learn, and its meaning use existing symbols (e.g. with TEX) than to create good, should be easy to remember, at least aer one has heard useful new symbols (e.g. with METAFONT). So TEX with or read an explanation once; i.e. a notation should be the character set oered by Computer Modern fonts (and mnemonic. the AMS fonts) shaped the typography of mathematics in A lot of mathematics is still (and will be) written by the past çý years. hand (e.g. in a mathematician’s research as the fastest way is situation only changed with Unicode mathemat- to denote his thoughts, on the blackboard, etc.). So a nota- ics: Unicode now oers mathematical symbols literally tion should be writable. In fact mathematical typography by the thousands. But it gives little explanation and little shows its close relation to handwriting in many places. But usage information; many symbols are described only by while written mathematics could always be explained by shape, not by meaning. For many Unicode mathematical the writer (e.g. by the teacher at the blackboard), printed symbols it is not clear how to use them, and in many cases mathematics has to speak for itself. So in some cases it it is not clear whether there are any competing or superior is desirable to go for greater dierentiation in print than notations. what is possible in handwriting. Math never seen 222 TUGboat, Volume 31 (2010), No. 2 A notation should also be pronounceable. Usually this is not a problem: for most notations there is a manner of speaking, although oen language-specic and oen not closely related to the notation (e.g. we call “ a ” the “absolute value of a”, and we would do so whatever the notation would be). But we’ ll see an example below where a missing manner of speaking was a problem. | | A new notation should be consistent with the general system of mathematical notation and similar to existing notations (e.g. for a symmetric relation one should choose a symmetric symbol, for a new kind of mapping one should choose some kind of arrow). In print, we can dierentiate more than in handwriting, but still it is oen preferable to stay close to existing notations. As a special case of similarity, there are many concepts in mathematics which are dual or complementary to each other, and such dual concepts should be given dual notations (e.g. and ; and ; and ; and ). Conversely, dual symbols should denote dual concepts. In some cases dual symbols work against mnemonics. For many students< it is> dicult∧ ∨ to∪ remember∩ ⊂ which⊃ is which, so one has to use an additional memory aid (e.g. to remember which one of the the logic symbols or denotes the “logical or”, one might learn that reminds of Latin “vel”, which means “or”). Of course a new notation should be distinct and un- ∧ ∨ ambiguous . Otherwise it will not be an improvement∨ upon existing notations. A notation should be adaptable, it should allow for manipulation. Also mathematical concepts are oen gen- eralized, and thus notation is oen stretched to more gen- Figure Ô: Robert Recorde, e Whetstone of Witte (London, eral cases. A good notation allows for that. Ô Þ). Recorde’s explanation for his symbol “ ” is given in A historical example is given by the competing no- the lines just above the display formulae. tations x of Newton and dx of Leibniz. While Newton’s notation was similar to existing notations and better tted = into the general system, the novel notation of Leibniz was ç Historical examples superior, as it was more versatile and allowed for manipu- ̇ To illustrate these quality criteria, I will give a few histor- lation and generalization. ical examples, some unsuccessful, some successful. e To give another example, the greatest common divi- historical information is mainly taken from [Ô]. sor of two integers a and b could be denoted as gcd a, b ; alternatively one might think of an inx notation, e.g. a b. ç.Ô Symbols for equality When applied to three arguments both notations still work: Our modern symbol for equality “ ” was introduced by gcd a, b, c and a b c . But one could also take the gcd ( ) Robert Recorde in Ô Þ in his book “e Whetstone of of all elements of a set S . With the rst notation, we can ⊤ Witte” (see gureÔ). Recorde explained his choice thus: write this as gcd S . Yet the alternative notation fails, it is “And to avoide the tediouse= repetition of these not( adaptable) enough.⊤ ⊤ And last on our list, a symbol should be available. woordes : is equalle to : I will sette as I doe oen in is is not really a criterion for quality, but rather for ac- woorke use, a paire of paralleles, or Gemowe lines ( ) ceptance. e best notation does not help much if other of one lengthe, thus: , bicause noe .ò. thynges, people are not able to use it. In former times, this mainly can be moare equalle.” meant availability at the printer’s oce — nowadays it (“Gemowe” means “twin”). is is quite a famous example, means availability in a font, then a clear and simple shape as it is one of the very few==== cases where an author not only which can be added to other fonts with ease, and of course introduced a new symbol, but explained why he chose its inclusion in Unicode mathematics. particular form. Johannes Kuster¨ TUGboat, Volume 31 (2010), No. 2 223 But in fact these two symbols (and a few competing symbols as well) struggled for supremacy throughout the ÔÞth century. Descartes was the more eminent mathematician, and with his important works his notation also spread. General adoption of “=” as the symbol for equality came only in the early Ôth century, mainly because Leibniz and Newton both used it. ç.ò Symbols of Benjamin Peirce In Ô À, Benjamin Peirce introduced the symbols “ ” and “ ” to denote the numbers ç.Ô¥Ô À... and ò.ÞÔò... (see gureç). To my knowledge, these were the rst signicant symbols of American origin. ᭘ ᭙ Figure ç: Benjamin Peirce’s symbols for the numbers ç.Ô¥Ô À... and ò.ÞÔò... (from J. D. Runkle’s Mathematical Monthly, Vol. I, No.

Load more